Goodhart's Law: When a Measure Becomes a Target
Goodhart's Law explained through its origin in monetary policy, Strathern's famous phrasing, the four flavors of Goodhart, surrogation, reward hacking, and how to defend against it.
June 2026
January 2026
Navigating AI Safety and Alignment Challenges
AI alignment problem: making AI do what we truly intend, not just literal instructions. Challenge is human values are complex and hard to specify...
March 2024
AI's Impact on Society: Benefits and Challenges
AI is transforming medicine, labor markets, and governance in real time. What do leading researchers actually think about the risks and benefits —...
June 2023
AI Career Paths: Opportunities and Transitions
AI career paths compared: ML engineer, AI researcher, AI product manager, AI safety, MLOps. Salaries by role and level, educational paths, and how...
February 2023
Defining AGI: Insights into Artificial General Intelligence
AGI refers to AI that matches or exceeds human cognitive abilities across all domains. Experts disagree sharply on timelines and what AGI would...
October 2022
What Is the Sycophancy Problem in AI and Why It Matters
AI sycophancy occurs when language models agree with users to seem helpful rather than telling the truth.
August 2022
Understanding the Principal Hierarchy Problem in AI Safety
The principal hierarchy problem is central to AI safety. Learn about value alignment, RLHF limits, reward hacking, constitutional AI, and why...